Optimizing Harvard's Proposed Open Access Self-Archiving Mandate

It is a great (and widespread) mistake to treat the problems of (1) journal overpricing, (2) publishing reform, and (3) copyright reform as if they were all the same thing as the problem of (4) research access. They are not. Open Access (OA) to research can be provided, quickly, easily, and directly, without first having to solve (1)-(3). And, once provided, OA can help pave the way toward solving (1)-(3). But conflate OA with (1)-(3) from the outset, and all you do is delay and handicap OA.

Harvard faculty are voting today on an Open Access (OA) Self-Archiving Mandate Proposal.

The Harvard proposal is to try the opyright-retention strategy: Retain copyright so faculty can (among other things) deposit their writings in Harvard's OA Institutional Repository.

Let me try to say why I think this is the wrong strategy, whereas something not so different from it would not only have much greater probability of success, but would serve as a model that would generalize much more readily to the worldwide academic community.

[Note: The first two concerns I raised yesterday have since proved unfounded: [I had not yet seen the exact wording of the motion when I first posted this on Tuesday.] The mandate -- successfully adopted Tuesday afternoon -- applies only to articles, and only to the final, refereed draft of those articles. But the 3rd and principle concern (about mandatory copyright retention) stands.

(1) Articles vs. Books. The objective is to make peer-reviewed research journal articles OA. That is OA's primary target content. The policy has to make a clear distinction between journal articles and books, otherwise it is doomed to fuzziness and failure. The time is ripe for making journal articles -- which are all, without exception, author give-aways, written only for scholarly usage and impact, not for sales royalty income -- Open Access, but it is not yet ripe for books in general (although there are already some exceptions, ready to do the same). Hence it would be a great and gratuitous handicap to try to apply OA policy today in a blanket way to articles and books alike, covering exceptions with an "opt-out" option instead of directly targeting the exception-free journal article literature exclusively.]

(2) Unrefereed Preprints vs. Peer-Reviewed Postprints. Again, the objective is to make published, peer-reviewed research journal articles ("postprints") OA. Papers are only peer-reviewed publications after they have been submitted, refereed, revised, and accepted for publication. Yet Harvard's proposed copyright retention policy targets the draft that has not yet been accepted for publication (the "preprint"): That means the unrefereed raw manuscript. Not only does this risk enshrining unrefereed, unpublished results in Harvard's OA IR, but it risks missing OA's target altogether, which is refereed postprints, not unrefereed preprints.

(3) Copyright Retention is Unnecessary for OA and Needlessly Handicaps Both the Probability of Adoption of the Policy and the Probability of Success If Adopted. Copyright retention is always welcome wherever it is desired and successfully negotiated by the author. But there is no need to require retention of copyright in order to provide OA.

Sixty-two percent of journals already officially endorse authors making their postprints OA immediately upon acceptance for publication by depositing them in their Institutional Repository, and a further 30% already endorse making preprints OA. That already covers 92% of Harvard's intended target. For the remaining 8% (and indeed for 38%, because OA's primary target is postprints, not just preprints), they too can be deposited immediately upon acceptance for publication, with access set as "Closed Access" instead of Open Access. To provide for worldwide research usage needs for such embargoed papers, both the EPrints and the DSpace IR software now have an "email eprint request" button that allows any would-be user who reaches a Closed Access postprint to paste in his email address and click, which sends an immediate automatized email request to the author, containing a URL, on which the author need merely click to have an eprint automatically emailed to the requester. (Mailing article reprints to requesters has been standard academic practice for decades and is merely made more powerful, immediate, ubiquitous and effective with the help of email, an IR, and the semi-automatic button; it likewise does not require permission or copyright retention.)

This means that it is already possible to adopt a universal, exception-free mandate to deposit all postprints immediately upon acceptance for publication, without the author's having to decide whether or not to deposit the unrefereed preprint and whether or not to retain copyright (hence whether or not to opt out).

This blanket mandate provides immediate OA to at least 62% of OA's target content, and almost-immediate, almost-OA to the rest. This not only provides for all immediate usage needs for 100% of research output, worldwide, but it will soon usher in the natural and well-deserved death of the remaining minority of access embargoes under the growing global pressure from OA's and almost-OA's increasingly palpable benefits to research and researchers. (With it will come copyright retention too, as a matter of course.) It is also a policy with no legal problems, no author risk, and hence no need for loopholes and opt-outs.

Needlessly requiring authors instead to deposit their unrefereed preprints and to commit themselves to retaining copyright today puts at risk both the consensus for adoption -- [note that this first worry is now moot insofar as adoption by Harvard itself is concerned, because the Harvard mandate was successfully adopted!, but it still stands for adoption by other universities] -- and, if adopted, the efficacy of the Harvard policy itself at risk, because of author resistance either to exposing unrefereed work publicly or to putting their work's acceptance and publication by their journal of choice at risk. It also opens up an opt-out loophole that is likely to reduce the policy compliance rate to minority levels for years, just as did NIH's initial, unsuccessful non-mandate (since upgraded to an immediate deposit mandate), with the needless loss of 3 more years of research usage and impact.

I strongly urge Harvard to reconsider, and to adopt the Immediate-Deposit/Optional-Access mandate (ID/OA) that is now being adopted by a growing number of universities and research funders worldwide, instead of the copyright-retention policy now being contemplated.

Stevan Harnad
American Scientist Open Access Forum